Exxon Valdez Oil Spill Restoration Project Annual Report

Juvenile Herring Growth and Habitats

Restoration Project 97320T Annual Report

This annual report has been prepared for peer review as part of the Exxon Valdez oil Spill Trustee Council restoration program for the purpose of assessing project progress. Peer review comments have not been addressed in this annual report.

Kevin D.E. Stokesbury Evelyn D. Brown

Robert J. Foy Jody Seitz

Brenda L. Norcross

Institute of Marine Science University of Alaska Fairbanks

Fairbanks, Alaska 99775-7220

April 1998

Juvenile Herring Growth and Habitats

Restoration Project 97320T Annual Report

Study history: Restoration Project 97320T is the core of the Herring Recruitment Dynamics Project, a multi-investigator ecosystem study and part of the Sound Ecosystem Assessment (SEA; PWSFERPG 1993) program in Prince William Sound (PWS). SEA was initiated because the lack of knowledge of the ecological processes affecting pink salmon and herring confounded the identification of damage caused by the Exxon Valdez oil spill. The PWS herring population crashed in 1993 possibly due to a viral infection (VHSV). This viral infection occurs more frequently in fish exposed to oil. Local residents, frustrated by the loss of valuable fisheries and the inability to accurately identify the causes, strongly voiced support for research. They formed a group, appealed to the EVOS Trustee Council, and as a result of their effort, SEA was created in 1994. Research on juvenile herring began in April 1995.

Abstract: The purpose of this project is to determine spatial distributions and habitats of age 0 to 2 year old Pacific herring (Clupea pallzsi). In 1997 we completed 5 die1 acoustic surveys, sampling Eaglek, Whale, Ziakof, and Simpson Bays in March, May, July, August and October. Aerial surveys were completed in June and July. As our field sampling effort decreased, our focus shifted to data analysis. The Prince William Sound Pacific herring population was made up of 3 size modes representing young of the year, juvenile (1 -2 year old) and adult (sexually mature) herring. Herring of each size mode were contagiously distributed on different spatial scales. Cohorts (agelsize) of herring were spatially segregated. Juvenile herring aggregated in Bays for the first 2 years of their life. Prey availability and herring prey selection varied spatially and temporally. These spatial distributions define the area where the physical and biological variables occur determining Pacific herring life history and population size in Prince William Sound.

Kev Words: Clupea pallasi, Pacific herring, juvenile, habitat, Prince William Sound, distribution

Proiect Data: Spatial distribution (acoustic and aerial), length frequencies, zooplankton community structure, diet compositions and near shore ichthyofauna structure are part of the SEA data base (Project 97320-5 Information Services and Modeling).

CITATION: Stokesbury, K.D.E., E.D. Brown, R.J. Foy, and B.L. Norcross. 1998. Juvenile Herring Growth and Habitats, Exxon Valdez Oil Spill Restoration Project Annual Report (Restoration Project 97320T), Institute of Marine Science, University of Alaska Fairbanks, Fairbanks, Alaska.

Table of Contents

................................................................. Executive Summary 4

.......................................................................... Introduction 5

............................................................................ Objectives 6

Methods .............................................................................. 7

............................................................ Results and Discussion 9

..................................................................... Literature Cited 12

....................................................... Table 1 . List of researchers 13

.............................................................................. Figure 1 14

.............................................................................. Figure 2 15

.......................................................................... Appendix I 16

Stokesbury. K . D . E., J . Kirsch. E . D . Brown. G . L . Thomas. B . L . Norcross . 1998 . Seasonal variability in Pacific herring (Clupea pallasi) and walleye pollock (Theragra chalcogramma) spatial distributions in Prince William Sound. Alaska .

........................................................................ Appendix I1 44

Brown. E . D . Preliminary Documentation of Temporal and Spatial Variability of Pacific Herring. Other Forage Fish. and Seabirds in Prince William Sound. Alaska

......................................................................... Apendix III 60

Brown. E . D . and G . A . Borstad . Progress Report on Aerial Survey Development

....................................................................... Appendix IV 79

Foy. R . J., and B . L . Norcross . 1998 . Spatial and temporal differences in the diet of juvenile Pacific herring (Clupea pallasi) in Prince William Sound, Alaska .

........................................................................ Appendix V 101

Seitz. J . Distribution of Herring and Other Forage Fish as Observed by Resource Users Restoration Project 97320T Supplement . Annual Report

Executive Summary

This project is a component of the Herring Recruitment group of SEA, initiated to provide information on the herring population in Prince William Sound and restoration measures required after the Exxon Valdez oil spill. The Herring Recruitment group is examining the physical and biological mechanisms affecting the survival of juvenile Pacific herring (Clupea pallasi) and providing indices of recruitment into the fishing stock. To do this a conceptual model addressing three objectives: 1 .) overwintering survival model, 2.) summer habitat model, 3.) monitoring strategy, has been created. The Growth and Habitat project is determining: 1 .) horizontal and vertical distributions of juvenile herring, using hydroacoustic and aerial surveys, and the underlying biological (predator distribution, prey distribution) and physical variables (oceanographic conditions, substrata) influencing these distributions, 2.) survival rate of juvenile herring based on densities and yearclass distribution, 3.) summer growth rates in different areas, 4.) habitat quality, based on oceanographic conditions, energetics, growth rates and prey availability, 5.) larval drift based on the SEA oceanographic model, 6.) an overwintering survival model.

In 1997 we completed 5 die1 surveys, sampling Eaglek, Whale, Ziakof, and Simpson Bays in March, May, July, August and October. During all but the latter survey four or five vessels were employed: an acoustic vessel, a trawler, a seiner, a processing boat which also supported the inshore fry skiff, and an oceanographic vessel. Aerial surveys were completed in June and July. One vessel was used during the October survey and focused on developing a cost-effective monitoring strategy combining hydroacoustic, zooplankton and fish collections with oceanographic measurements. As our field sampling effort decreased, our focus shifted to data analysis. Presently, the majority of acoustic survey data has been combined with fish collection data to extrapolate estimates of spatial distributions for pollock and herring. Diet composition of juvenile herring and prey availability within the four bays has been determined for 1995 and 1996, and work continues on the 1997 data.

The Prince William Sound Pacific herring population was made up of 3 size modes representing young of the year. juvenile (1-2 year old) and adult (sexually mature) herring. Herring of each size mode were contagiously distributed on different spatial scales. Cohorts (agelsize) of herring were spatially segregated. Juvenile herring aggregated in Bays for the first two years of their life. They congregated in surface waters during June and July. Prey availability and herring prey selection varied spatially and temporally.

These spatial distributions define the area where the physical and biological variables occur determining Pacific herring life history and population size in Prince William Sound. The lack of information of these variables confounded the identification of damage caused by the Exxon Vclldez oil spill and estimates of the population's recovery time. Our future work will focus on determining these variables.

Introduction

The purpose of this project is to determine spatial distributions and habitats of age 0 to 2 year old Pacific herring (Clupea pallasi). It is linked to the Herring Recruitment Dynamic subgroup of SEA and provides data for the three objectives (1. overwintering survival model, 2. summer habitat model, 3. monitoring strategy) which will determine the physical and biological mechanisms influencing the recovery of Pacific herring. Pacific herring is listed as "not recovered" in the "Resources and Services Injured by the Spill" Exxon Valdez Oil Spill Restoration Plan.

The Herring Recruitment Model is being developed as the integration of several submodels, each of which focuses on a stage in the early life history of Pacific herring (Clupea pallasi). We hypothesize that, like other clupeoids, year-class strength of Pacific herring in Prince William Sound (PWS) is determined during its early life history. All field work, laboratory experiments, and data analysis for all components of SEA relate to one or more of these submodels. Two major SEA hypotheses are the focus of these submodels and will be linked within the overall Herring Recruitment Model (Figure 1). The first is the Herring Overwinter Hypothesis which states that survival of herring through their first winter is critical to year-class strength and is dependent upon their condition when they enter winter. We will test this hypothesis by examining distribution and condition of herring in the fall, winter and spring. We expect to see changes in condition indices related to the physical and biological variables of different geographic locations. A bioenergetic model, combining SEA field and laboratory observations together with energetic information from Atlantic herring studies, is being constructed to predict overwinter survival for recruiting herring. In support of the herring Overwinter Hypothesis we will examine how the LakeRiver hypothesis applies to transport and distribution of herring at the larval stage. We will employ larval drift simulations, using the Circulation and Transport Models for PWS being formulated by Mooers and Wang as part of the Ocean Dynamics Model, to determine the expected drift of larval herring within PWS and determine how that affects the distribution of summer juvenile nursery areas. We expect to examine various drift patterns in response to simulated lake (i.e. retention), river (i.e. rapid movement through the sound), and combinations of varying amounts of "lake" and "river" in accordance with the recent evolution of the lakelriver hypothesis. The larval drift synthesis is a tool, which will link the Summer Habitat Model, which examines location and characteristics of summer nurseries utilized by juvenile herring, with the Overwintering Survival Model. The Summer Habitat Model will determine the survival and growth rates of juvenile herring and the quality of nursery areas by examining changes in herring distribution, density, length, weight, energy ( k ~ ~ . ' ) , interspecific biological variables (prey abundance, predation) and physical variables (oceanographic conditions, bathymetry). These data will define the conditions of herring entering into the Overwintering Survival Model.

This project is a component of the SEA project; Dr. T. Cooney chief scientist. Within SEA, coordination exists between projects linking physical and biological data. Multiple authors on proposed publications reflect this integration. In addition, this project coordinates with

the APEX and NVP ecosystem projects via field logistics (vessels, equipment and samples), shared data (catch, aerial survey data, and acoustics results), and joint publications. Coordination with these groups increased during FY98 and we expect this to continue in FY99.

Objectives

The research objectives of this project are:

1. Develop an Overwinter Survival Model for juvenile herring.

2. Develop a Summer Habitat Model for juvenile herring.

3. Develop a Monitoring Strategy for juvenile herring.

For the Overwinter Survival Model:

Describe overwinter distribution, size, condition, energy needs, and relative abundance of juvenile herring, physical and biologic characteristics of herring nursery areas and overwintering bioenergetics.

Tasks:

1. Collect data on the whole body energy content of age 0 and 1 herring in the late fall and winter.

2. Determine changes in bioenergetics over the winter season using time sequence (monthly) sampling of juvenile herring from two or more index sites in 1996-97 and 1997-98.

3. Examine stomach contents of over-wintering recruits and make energetic estimates for consumption during the winter of 1996-97 and 1997-98.

4. Determine the energy need of fasting herring in the laboratory.

5. Develop a model to predict winter survivorship using field and laboratory measurements of over-winter energy needs and literature values for Atlantic herring.

6. Describe spring, pre-bloom biological and habitat conditions as an endpoint of Overwintering Survival Model and beginning of second year Summer Habitat Model.

7. Compile historic biological and physical data for the purpose of model verification

For Summer Habitat Model:

Describe summer and fall distribution, size, condition and relative abundance of juvenile herring (biological data), and physical and biological characteristics of herring nursery areas (habitat data) to evaluate quality of summer growth of herring and as initial conditions for the Overwintering Survival Model.

Tasks:

1. Use Circulation and Transport Models (Ocean Dynamics Model) to simulate drift of larval herring and distribution to summer nursery areas.

2. Determine distribution of juvenile herring during the spring, summer and fall using broad scale surveys that include simultaneous overflights, acoustics and net collections.

3. Determine physical (salinity, temperature, depth, currents, light levels, bathymetry) and biological (zooplankton, competitors) parameters which determine good vs. bad nursery areas measured by condition of herring (length, weight, age, growth rates, stomach contents, energetic condition and stable isotopes).

4. Develop maps of key habitats (nursery areas) for juvenile herring within PWS.

5 . Describe the retention characteristics of herring nursery areas using information from the larval drift simulations, physical oceanographic measurements and biological data (spatial distributions, isotopes, and growth rates) indicating immigration or emigration.

6 . Develop maps of possible retention areas with different historical spawning sites and transport conditions.

7. Compare historic distributions reported by local and traditional knowledge with distributions described by this study

For Monitoring Strategy:

Tasks:

1. Identify key index sites and develop monitoring techniques by relating aerial, acoustic and net sampling data during summer surveys to condition of juvenile herring.

Methods

To address the above objectives and tasks, we have formulated our approach into two component models. each with several subcomponents. These models and subcomponents are described in chronological order of herring life history (Figure 1).

The first subcomponent is embryo survival. This component is not a SEA program, but rather projects funded by EVOS outside of SEA. For the starting point of our Summer Habitat Model, we intend to combine the results of 1) the ADF&G spawn deposition survey, 2) the Haldorson, Quinn and Rooper egg loss model which predicts losses due to physical factors and predation, 3) estimates of baseline egg mortality, and 4) estimates of baseline levels of viable hatch (Hose et al. 1996; Kocan et al. 1996). From this we will know the location of spawning of herring, an estimate of the amount of spawn, and the expected percentage of viable larvae produced.

The output of that subcomponent is the input into the Larval Drift Model (Figure 1). We will examine the direction of transport without incorporating the population size component. We will run the Ocean Circulation and Transport Model with input at the locations of herring spawning and test observed distribution of particles. Distribution predicted by this subcomponent will be verified by the distribution of age-0 herring during the summer. We will compare the Larval Drift Model results to the transport and retention of larval Atlantic herring (Clupea harengus) in North Atlantic (Graham and Davis 1971, Graham and Townsend 1985, Sinclair and Iles 1985, Sinclair 1988). We will also use 1989 as a test case. By inputting location of spawning and physical conditions which we know occurred in 1989, we can test the model against the offshore distribution of larvae observed in May, June and July 1989 (Norcross and Frandsen 1996) and the nearshore distribution observed in May 1989 (McGurk 1990). We will also use spawning location information from 1995, 1996 and 1997 correlated with the distribution of larvae and the distribution of herring observed from the aerial and acoustic surveys. This simulation will be an iterative process.

The output of the larval drift simulation is the input for the Summer Habitat Model (Figure 1). From October 1995 to March 1998, acoustic and aerial surveys were conducted and these data will be processed, analyzed, interpreted and combined to determine herring nurseries. The broadscale distribution of age-0 herring was observed during October 1995, March and July 1996. These surveys covered most of PWS and adjacent waters to Resurrection Bay. The Sound is very large and resources were limited so the survey focused on regions where fishermen had observed juveniles and where earlier ADF&G surveys indicated high densities of herring (

fish distribution and oceanographic conditions are strongly effected on a short temporal scale by the tidal cycle and the diel (daylnight) cycle.

We have addressed these limitations by employing a factorial design, based on Green's (1979) principles of sampling to derive the survival rate of juvenile herring from density changes using a life table. Densities must be estimated precisely and accurately on both spatial and temporal scales. In this factorial design each spatial replicate (bay) has 3 temporal replicate samples within 24 hours, allowing us to estimate the variability in densities caused by tidal and diel cycles and allow accurate measurements of the oceanographic conditions of each bay (Gunderson, 1993 ) (Figure 2). This design allows an overall estimate of changes in survival rates of Prince William Sound juvenile herring and comparisons between and within bays on different spatial and temporal scales, i.e. 24 hours, monthly, annually. The four bays are, Eaglek, Whale, Ziakof and Simpson. These bays were selected because:

1 . herring overwinter in bays

2. spatially segregated; North South, East and West.

3. located at a distinct position along the prevailing PWS current, relating directly to the lakelriver hypothesis (Cooney 1995: Chapter 7 Fig. 1 1 - 17; and Chapter 9).

4. strong evidence that herring spawnlrecruitment in each of these bays

Each bay was surveyed three times in a 24 hour period using sidescan sonar (Figure 2). Net collections of herring were coupled with acoustics estimates of horizontal and vertical distribution and abundance, and aerial estimates of horizontal distribution. These net collections are used to ground-truth both acoustic and aerial estimates for species size and composition. Subsamples of herring were retained and later evaluated for size, age, stomach contents, condition (energetics and standard fisheries age-weight-length (AWL)), and stable isotopes (trophic analysis). Simultaneous with net collections for fish were vertical plankton tows to estimate availability of food for planktivorous herring. Oceanographic parameters collected include salinity and temperature at depth (CTD), estimates of current structure (ADCP), light levels and bathymetry at location. The main effort in 1998 will be to process, analyze and interpret these data.

Results and Discussion

Larval Drift Model

The Ocean Circulation and Transport Model is being developed and the first results are being published in the manuscript:

Mooers, C.N.K. and J. Wang. 1998. On the development of a three-dimensional circulation model for Prince William Sound, Alaska. Continental Shelf Research 17:OOO-000 in press.

This model will be the basis for the Larval Drift Model, which is presently being developed.

Summer Habitat Model

In this component we are determining the biological and physical variables influencing the spatial and temporal distribution of Pacific herring (Clupea pallasi) in Prince William Sound. This is a combined effort with support for acoustics and oceanography from PWSSC and technical support from Alaska Department of Fish and Game in Cordova.

The first manuscripts from this work are:

Stokesbury, K. D. E., J. Kirsch, E. D. Brown, G. L. Thomas, B. L. Norcross. 1998. Seasonal variability in Pacific herring (Clupea pallasi) and walleye pollock (Theragra chalcogramma) spatial distributions in Prince William Sound, Alaska. Appendix I

Brown, E.D. , S. Vaughan, and B.L. Norcross. In prep. Annual and seasonal spatial variability of herring, other forage fish, and seabirds in relation to oceanographic regimes in Prince William Sound, Alaska in Ecosystem Considerations in Fisheries Management, Lowell-Wakefield Symposium. A preliminary look at data to be included in this publication is attached in Appendix 11.

Brown, E.D., G.A. Borstad, and B.L. Norcross. In prep. Assessment of forage fish distribution, relative abundance, and ecology using aerial surveys: survey design and methodology. Ecological Applications draft. Chapter 1 1, Appendix I, SEA 1996 Annual Report. Appendix 111 is a progress report on this work.

Foy, R. J., and B. L. Norcross. 1998. Spatial and Temporal Differences in the Diet of Juvenile Pacific Herring (Clupea pallasi) in Prince William Sound, Alaska. Appendix IV

Overwintering Survival Model

The Overwintering Survival Model evaluates distribution and condition of age 0 and 1 herring as they enter, pass through, and complete the winter. The objective of this sampling is to determine change in condition of herring over the course of winter in concert with the hypothesis that herring which enter winter in poor condition due to " b a d nursery habitats will not survive winter, while those from "good" habitats will successfully survive winter.

Dr. A.J. Paul is leading this effort, he and his co-authors have submitted the following manuscripts:

Paul, A.J., J.M. Paul, and E.D. Brown. 1997. Fall and spring somatic energy content for Alaskan Pacific herring (Clupea pallasi Valenciennes 1847) relative to age, size and sex. Journal of Experimental Marine Biology and Ecology. 000:OOO-000 in press.

Paul A.J., and J. M. Paul. 1997) Comparisons of whole body energy content of captive fasting age zero Alaskan Pacific herring (Clupea pallasi Valenciennes) and cohorts over-wintering in nature. Journal of Experimental Marine Biology and Ecology. 000:OOO-000 in press.

Refer to Dr. Paul's sections in this Annual report for more details on these manuscripts.

Alaska Predator Ecosystem Experiment (APEX) Project Support

Juvenile herring were determined to be an important forage fish in PWS (Haldorson et al. 1996). Thus there is considerable overlap between research conducted by SEA and APEX. In 1997, SEA herring researchers were requested to cooperate and share data with researchers within APEX. Overflights were coordinated with APEX acoustic surveys and ongoing sea bird research. The aerial database was shared with APEX to enhance modeling efforts linking fish distribution to bird foraging behavior and reproductive effort. In addition, SEA aerial and net catch data concerning jellyfish was shared with the APEX project for analysis and publication. We expect these cooperative efforts to continue as both programs move toward synthesis.

Historic Data Summary

An effort to capture local knowledge about herring and other forage fish continued, for its second year, as a supplement to this project. Presently thirty-nine individuals have been interviewed. Several observations were consistent among interviewees: juvenile herring are most abundance at the heads of bays in the summer, juvenile herring have a different distribution than adults, juvenile herring schools are smaller but more abundant that adult schools. These anecdotal observations are in general agreement with the results of this study. A separate annual report has been prepared for 97320T supplement Appendix V.

Literature Cited

Graham, J.J. and C.E. Davis. 1971. Estimates of mortality and year class strength of larval herring in western Maine, 1964-67. ICES Rapp. R.-V. 160: 147-152.

Graham, J.J. and D.W. Townsend. 1985. Mortality, growth, and transport of larval Atlantic herring Clupea harengus in Maine Coastal waters. Trans. Am. Fish. Soc. 114: 490- 498.

Gunderson R.D. 1993. Surveys of Fisheries Resources. John Wilet & Sons, Inc., New York

Haldorson, L., T. Shirley, K. Coyle, and R. Thorne. 1996. Alaska Predator Ecosystem Experiment Forage Species Studies in Prince William Sound Project 163A, 1996 Annual Report prepared for Exxon Valdez Oil Spill Trustee Council, Anchorage, Alaska. 93 pp.

Hose, J.E., M.D. McGurk, G.D. Marty, D.E. Hinton, E.D. Brown, and T.T. Baker. 1996. Sublethal effects of the Exxon Valdez oil spill on herring embryos and larvae: morphologic, cytogenetic and histopathologic assessments, 1989- 199 1. Can. J. Fish. Aquat. Sci. 53: 2355-2365

Kocan, R.M., J.E. Hose, E.D. Brown, T.T. Baker. 1996. Herring embryo (Clupea pallasi) sensitivity to Prudhoe Bay petroleum hydrocarbons: laboratory evaluation and in situ exposure at oiled and unoiled sites in Prince William Sound. Can. J. Fish. Aquat. Sci. 53: 2366-2375

McGurk, M.D. 1990. Early life history of Pacific herring: Prince William Sound herring larvae survey. Final report to NOAA-NOS, Contract 50ABNC-7-00141, Anchorage, AK.

Norcross, B.L. and M. Frandsen. 1996. Distribution and abundance of larval fishes in Prince William Sound, Alaska during 1989 after the Exxon Valdez oil spill. In S.D. Rice, R.B. Spies, D.A. Wolfe and B.A. Wright (eds.). Exxon Valdez Oil Spill Symposium Proceedings. Am. Fish. Soc. Symp.

Sinclair, M. 1988. Marine Populations. Washing Sea Grant Program, Seattle, WA, 252p.

Sinclair, M. and T.D. Iles. 1985. Atlantic herring (Clupea Izarengus) distributions in the Gulf of Maine-Scotian Shelf area in relation to oceanographic features. Can. J. Fish. Aquat. Sci. 41 : 1055-1065.

Table 1. List of researchers we collected samples for during the SEA Herring cruises.

1. Kathy Frost, ADF&G Fairbanks, AK.; Marine Mammal Ecosystem. Needed various size fish of any species.

2. Jeff Short, Auke Bay, Juneau AK.; Needed herring and pollock.

3. Molly Sturdavent, Auke Bay, Juneau AK.; Needed capelin, sandlance, eulachon and pollock.

4. Tom Kline, PWSSC, Cordova AK.; Isotopes

5. John Piatt, NBS, Anchorage AK.; Needed juvenile herring and pollock

6. A.J. Paul, Seward Marine Center, Energetics, herring and pollock.

7. James Raymond, Univ. of Nevada; Needed blood and liver samples from herring in the Gravina or Montague area.

8. Steve Moffitt and John Wilcox, ADF&G Cordova. Herring AWL.

9. Richard Kocan, UW; Disease; 60 juvenile herring from 5 sites, would like heart, liver and spleen removed, put in tubes, and kept cool.

10. Gary Marty, UC-Davis; Disease; Looking at herring from the Montague Is. area

L;~rv:~e Juverule (3-6 mon.)

M:trcl~-May July October

(bioenergetics)

oceanography oceanograph

(consumption)

(sizdgrowth

(survival) overwinterir~g

August

labitat

(I,iwnergetics)

spaw11 energetics

Ispawn deposition 1% recruits into (survival) Idensityldid. (egg loss model) adult schools

v >2 yr old J u v l ( I yr, 3 o n . G a - 3

mortalities

Figure 1 . Hcrring Rccruitnlent Model

SEA Herring Survival-Growth Sampling Design

7 day survey of Prince William Sound

Oct. 95 to Mar 98

Y2 age 1 herring X2 salinity Y 3 age 2 herring X3 temperature Y4 mature herring X4 bathymetry YS larval herring X5 light intensity Y5 juvenile pollock X6 freshwater input Y6 adult pollock X7 tidal cycle Y7 other species X8 light Y8 zooplankton Y9 jellyfish Y 1 0 predators

800 1600 2400

Y = dependent or independent variable X = independent variable

Eaglek Bay Whale Bay Zaikof Bay

Simpson Bay

Figure 2. 3-bay die1 sampling design.

AWL, energetics, aerial survey AWL, energetics, aerial survey AWL, energetics, aerial survey AWL, energetics, aerial survey

Sample Unit

Biological Factors Environmental factors

Y1 YOY herring X1 SimpsonfHunter parameter

Appendix I

Stokesbury, K. D. E., J. Kirsch, E. D. Brown, G. L. Thomas, B. L. Norcross. 1998. Seasonal variability in Pacific herring (Clupea pallasi) and walleye pollock (Theragra

chalcogramma) spatial distributions in Prince William Sound, Alaska. For submittion to Marine Ecology Progress Series.

Seasonal variability in Pacific herring (Clupea pallasi) and walleye pollock (Theragra chalcogramma) spatial distributions in Prince William Sound, Alaska.

Kevin D. E. ~tokesbury'

I Institute of Marine Science, University of Alaska-Fairbanks, Fairbanks, Alaska, 99775-7220

Telephone: (907) 474-5 184; Fax: (907) 474-1 943 email: kstokes @ ims.alsaka.edu

Jay ~ i r s c h *

'prince William Sound Science Center P.O. Box 705, Cordova, AK, 99574

Evelyn D. ~ r o w n '

Gary L. ~homas '

Brenda L. orc cross'

ABSTRACT: Pacific herring, Clupea pallasi, and walleye pollock, Theranra chalco~ramma, - spatial distributions were determined using acoustic surveys, with supporting net collections, in October 1995, March and July 1996 in Prince William Sound, Alaska. Of the 97 species of fish and macroinvertebrates collected, Pacific herring (65.0%) dominated the nearshore ichthyofauna, followed by walleye pollock (19.2%), Pacific sand lance (Ammodytes hexapterus; 2.6%), and capelin (Mallotus villosus; 1.9%). The Pacific herring population size structure was trimodal representing age 0. 1 to 2 year old, and adult fishes. The walleye pollock population size structure was bimodal representing age 0 and adult fishes. Large scale distributions of Pacific herring and walleye pollock were contagious, with aggregations occurring in the east-northeast and the west- southwest areas of the Sound. Pacific herring occupied the upper 30 m of the water column while walleye pollock were usually located near the bottom. Bays appeared to be nursery areas as age 0 and 1 to 2 Pacific herring were aggregated within them during all surveys. After their second winter juvenile herring joined the adult schools, leaving the bays at approximately the same time that new recruits enter the bays. Adult Pacific herring migrated seasonally, overwintering in Zaikof Bay and spending the summer in the west- southwest portion of the Sound.

Key words: Pacific herring, Clupea pallasi, walleye pollock, Therapra chalcogramma, acoustics, spatial distribution, nursery area, migration, size distribution, oil spill

Introduction An organism's life history is it's response to physical and biological variables

allowing it to persist in a specific geographic area over time (Sinclair 1988). In Prince William Sound, Alaska, the Pacific herring (Clupea pallasi Valenciennes, 1847) population crashed in 1993 (Funk 1994, Paine et al. 1996). Prince William Sound contains a biologically rich, high latitude ecosystem and little is known of the physical and biological variables influencing Pacific herring life history (Paine et al. 1996). This lack of information confounded the identification of damage caused by the 24 March 1989 Exxon Valdez oil spill (36,000 metric tons of north Slope crude oil effecting 900 km of coast line in PWS), commercial fisheries (bait, sac-roe, and roe on kelp) and estimates of the population's recovery time (Paine et al. 1996, Spies et al. 1996). This is the second pulse perturbation effecting Prince William Sound's marine community in 34 years, the 27 March 1964 earthquake raised sections of the coast line by as much as 3 m (Hansen & Eckel 1971). A third pulse perturbation may be occurring with this year's large El Nino. Fish communities in other highly perturbed systems have shifted their species composition and abundance to new equilibrium levels, for example large sand eel, pollock and eel pout populations replaced herring and mackerel populations in the North Sea due to high fishing mortality on the latter species (Andersen & Ursin 1978, Auster 1988). Walleye pollock (Theragra chalcogramma Pallas, 18 14) may be a competitor and predator of Pacific herring and may have filled the vacancy in the nearshore ecosystem that occurred due to the decrease in herring density (Willette et al. 1997).

Pacific herring usually begin to spawn in their third year when they have reached a size of about 185 mm and a weight of 95 g (Robinson 1988). Females can produce as many as 40,000 eggs each year until they reach an age of about 15 years (Robinson 1988). Pacific herring deposit their eggs in mid-April in the nearshore low intertidal or subtidal zone, primarily on marine vegetation (Wespestad & Moksness 1990, Brown et al. 1996). In early May, after approximately 2 weeks, the eggs hatch into larval herring. They metamorphose from the larval to juvenile form when they reach a size of 25 mm to 30 mm, which can take from 4 to 10 weeks (Wespestad & Moksness 1990). During this time larvae are transported away from the spawning areas, although studies in British Columbia, Canada, have found significant densities remaining nearshore (Robinson 1988).

Walleye pollock, (Theragra chalcogramma) are an important commercial species in Alaskan waters and a primary forage fish for sea birds, marine mammals and fish (Clausen 1983, Hatch & Sanger 1992, Livingston 1993). Walleye pollock congregate and spawn in deep water in late March and April and the larvae occupy the upper 50 m of the water column in late May (Hinckley et al. 1991, Kendall et al. 1996). Walleye pollock metamorphose into juveniles in August and September (Hinckley et al. 1991, Kendall et al. 1996).

Acoustic estimates of fish abundance are frequently used for stock assessment and fisheries management but infrequently for ecological and life history studies (Thorne 1983a, Thorne 1983b, MacLennan & Simmonds 199 1 , Thomas 1992, Gunderson 1993. Misund 1997). Acoustic surveys estimating Pacific herring fishing stocks have been conducted from Alaska to California. but relatively few have been conducted in Prince William Sound (Thorne 1977a, Thorne 1977b, Trumble et al. 1982, Thorne et al. 1983, Thorne & Thomas 1990).

We examined seasonal variability in Pacific herring, Clupea ~allasi, and walleye pollock, Theragra chalcogramma, spatial distributions in Prince William Sound, Alaska. We hypothesized that: 1. Pacific herring and walleye pollock were contagiously distributed and this distribution varied seasonally; 2. Pacific herring and walleye pollock occurred in different areas of the water column; 3. cohorts (agelsize) of Pacific herring were spatially segregated; 4. juvenile Pacific herring aggregated in bays rather than passages or along open coastline. To test these hypotheses we determined the spatial distributions of Pacific herring age cohorts and walleye pollock observed during three acoustic surveys of the Prince William Sound coastline in October 1995, March and July 1996. Material and Methods.

Prince William Sound is a large body of water separated from the Gulf of Alaska by a series of mountainous islands and deep passages (Fig. 1). The rocky coastline is highly irregular with numerous islands, passages, bays and deep fjords. The Sound has a semi- diurnal tide with a maximum range 4.4 m during this study.

The Prince William Sound coastline was acoustically surveyed in October 1 995, March and July 1996 (Fig. 1). Five vessels were used during each 10 day survey (12 hours per day); three commercial fishing vessels ( ~ 1 6 . 8 m) which deployed the acoustic and oceanographic equipment and fished the seines, a trawler (Alaska Department of Fish and Game R N Pandalus, =20 m trawler), a cruise vessel (=25 m) where the samples were processed. Surveys were conducted during the night (2000 to 0800 h) in October 1995 and March 1996 and in daylight (0800 to 2000 h) during July 1996.

The acoustic vessel followed a zig-zag pattern along the shore to a distances of = 1 km at a speed of 14 to 17 krn h-'. The vessels sonar (50 KHz, 46") was used to locate schools along the transect. When a school of fish was encountered the acoustic vessel slowed to 9 to 11 km h-' and completed a series of parallel transects perpendicular to shore using a 120 kHz BioSonics 101 echosounder with a preamplifier dual-beam transducer. The transducer was mounted on a BioSonics 1.2 m BioFin and towed =1 m under the water surface. The acoustic signals were processed in real-time using the BioSonics ESP 221 Echo square integration software and ESP 28 1 Dual beam software, and stored on digital audio tape (Thorne 1983a, Thorne 1983b, MacLennan & Simmonds 1991). The acoustic system was calibrated both with a hydrophone and a standard target. Parameters of the acoustic system during the surveys were: source level = +225.023 dB; receiver gain = - 159.28 dB; transducer directivity = 0.00107 and pulse duration = 0.4 ms. The signal to noise ratio exceeded that required to estimate the minium fish densities. The acoustic survey vessel used a GPS navigational system, and coordinates were imported by the ESP software and C-MAP software to plot the survey tracks.

Once the acoustic vessel surveyed a fish school, one of three fishing vessel sampled it to determine species composition and size structure. Fish were sampled using a modified bottom trawl for deep water targets (1.52 x 2.13 m Nor'Eastern Astoria V trawl doors, head rope 2 1.3 m, foot rope 29.0 m, estimated 3 x 20.0 m mouth, 10.2 cm mesh wings, 8.9 cm middle and a 32.0 rnm cod end liner), an anchovy seine for surface schools (250.0 x 34.0 m and 20.0 m, 25.0 mm stretch mesh) or, in shallow water, a small salmon fry seine (50.0 x 8.0 m, 3.0 mm stretch mesh deployed from a 6m skiff equipped with a 70 horsepower engine). Each collection was speciated and 1000 herring, or other dominant fish species in the catch, were randomly subsampled. Fork length (mm) and weight (g) were recorded

from 450 herring and remaining fish were measured to fork length. Seine collections were often very large, therefore once the subsample had been collected the remaining fish were released unharmed.

A length dependent scaling constant was used to convert the reflected acoustic energy into a biomass estimate:

TS re:w (dB re:kg) = -6.0 logx - 24.2 dB where x is the mean fork length (cm) of the fish caught in the area (Thorne 1977b, Thorne 1983a, Thorne 1983b, Thome& Thomas 1990). This equation differs from the more standard regression equation calculated by Foote (1987) as it derives the target strength as a proportion of weight. It was developed for echo integration primarily using Pacific herring surveys from Alaska and Puget Sound (Thorne 1983a). However, Thorne's equation was unsuitable for estimating walleye pollock target strength, due to the fish's morphology, therefore the more standard equation:

TS = 20 logx - 66.0 dB was used (Foote & Traynor 1988, MacLennan & Simmonds 1991).

Echo integration measurements were converted into data cells with 120 to 40 m, 40 m or 20 m lengths (estimated from 2 ping s-I with a surveying speed of 2.5 to 3.0 m s-' equals 60 or 32, 32, 16 pings cell-' respectively) and 1 m widths and depths, during the October 1995, March 1996, and July 1996 surveys, respectively. Latitude and longitude were recorded simultaneously with each data cell from the GPS. The acoustic files were transferred to a UNIX work station for batch processing. Software created in the Interactive Data Language (DL) was used to apply the acoustic calibrations and enabled interactive image editing to remove untracted bottom and other non-biological scatter.

Species proportion and size modes per species were determined from the fish collections. The species proportions, based on the number individuals per fish species in the random subsample, were converted to biomass using lengthlweight regressions. Using these proportions the echo integration densities (kg m-3) were converted into number of Pacific herring per size mode, or number of Walleye pollock. Walleye pollock were not divided into size modes because the standard deviations of the mean fork lengths for individual collections indicated that aggregations were primarily a single size mode. Based on frequency distributions of the data we assumed that cells containing the equivalent of ~ 0 . 5 fish m-' were probably zooplankton, so they were removed from the data set (MacLennan & Simmonds 199 1, Gunderson 1993). Fish located near the bottom were difficult to distinguish acoustically, so if the signal appeared to be corrupted the bottom 5 m were removed. Visual examination of the echograms and fish collections agreed with these assumptions.

The hypotheses that Pacific herring and walleye pollock were contagiously distributed, were spatially segregated, and that Pacific herring cohorts (agelsize) were spatially segregated were tested by determining the large scale spatial distributions of herring size modes, calculated from fork length frequencies, using circular statistics (Batschelet 198 1 ). The angle (0' = true north) for each data cell, within each herring size mode, was determined from an origin in the center of Prince William Sound (60" 60.00', 146" 90.00'; Fig. 1 ). These angles represent distributions along the transect line and are influenced by inequalities in shore line distance and sampling bias. These angle frequency distributions were compared to random distributions and to the distributions of other

herring size modes along the same transect using a chi-squared test at the 5 % levels of significance (Batschelet 1981). Expected values were grouped according to Cochran's rule, which states that 0.009'. Visual examination of the echograms and plots of cells along transects supported this assumption. Bays were statistically defined from passages or open coast line by creating a parameter that is the sum of the three nearest shore distances (x3NSD), each separated by 90". This measurement was calculated at 12 bays and along 2 passages, 26 inside bays and 17 outside bays, to verify that it accurately defined bays, passages, and open coast line. The C3NSD for each herring school within each size mode was calculated and compared to the same measurements from randomly selected points along the transect for the three surveys using a chi-squared test. Expected values were grouped according to Cochran's rule (Sokal & Rohlf 198 1). This technique corrected sampling and shoreline structure biases, for example 80 % of the shoreline may have been in bays or we may have surveyed bays more than open coastline or passages. Further to determine if there were physical differences in water mass and larval fish community structure at these locations vertical water profiles measuring temperature and salinity at 1 m intervals, using a SeaBird instrument, and two tucker trawl samples (1 m' mouth, 505p mesh nets) were collected at each of these sites. Tucker trawl samples were preserved in 10% formalin for >48 h, then transferred to 50% isopropyl acholol and sorted to species and measured in the laboratory. Results.

Ninety-seven species of fish and macroinvertebrates were collected during October 1995, March and July 1996 surveys. Pacific herring (Clupea pallasi; 65.0%) dominated the ichthyofauna; followed by walleye pollock (Thera~ra chalcogramma; 19.2%), Pacific sand lance (Ammodvtes hexapterus Pallas, 18 14; 2.6%), and capelin (Mallotus villosus Muller 1776; 2.0%) (Table 1). Pacific herring and walleye pollock were collected in all fishing gear during each survey while the majority of capelin and Pacific sand lance were collected in seines during March and July 1996, respectively (Table 1).

The Pacific herring population consisted of three size modes representing age 0, 1 - 2 year old, and adult fishes (Fig. 2). In October 1995 the first mode (0-1 10 mm) were the 5 months old 1995 cohort, the second mode (1 1 1 - 180 mm) were the 1 year 5 month old 1994 cohort and the third mode are mature adults. The March 1996 the first mode (0- 120 mm) were the 10 months old 1995, the second mode ( 1 2 1 - 180 mm) were the 1 year 2 month old 1994 cohort and the third mode are mature adults. In July 1996 the first mode (0-80 mm) were the 3 months old 1996 cohort, the second mode (8 1 - 160 mm) were the 1 year 3 month old 1995 cohort and the 1994 cohort moved into the adult mode. The walleye pollock population was bimodal representing age 0 and adult fishes, in October 1995 and March 1996 while only age 0 fish were collected during July 1996 (Fig. 3).

Pacific herring and walleye pollock were contagiously distributed within Prince William Sound (Fig. 4). The general distribution of both species was bimodal with aggregations occurring in the east-northeast and the west-southwest (Fig. 1 and 4).

Location of age 0 and 1 - 2 herring and walleye pollock aggregations were relatively similar within this bimodal distribution during each survey (Fig. 4). Adult herring appeared to be aggregated in the southwest (210'-240') in October 1995, in the south (180') and east (90") in March 1996, and west (240'-270') in July 1996 (Fig. 4).

The distributions of the three size modes of Pacific herring differed from each other and from walleye pollock in October 1995, March, and July 1996 (Table 2; Fig. 4). Age 0 herring had a similar distribution to age 1 - 2 herring in October 1995 but differed in March and July 1996 (Table 2; Fig. 4). Age 0 herring distribution always differed from adult distribution (Table 2; Fig. 4). Age 1 - 2 and adult herring distributions were similar in March 1996 but differed during the other two surveys (Table 2; Fig. 4). Walleye pollock distribution differed from all three herring size mode distributions combined and from each size mode except for Age 0 herring in March 1996 (Table 2; Fig. 4).

Pacific herring schools had the lowest densities for all size modes in March 1996 and the highest densities in July 1996 (Table 3). Pacific herring schools had the highest average number of data cells in March 1996 (Table 3). Walleye pollock schools had low densities in October 1995 and March 1996 compared to very dense schools in July 1996 (Table 3).

Pacific herring were primarily distributed top half of the water column. Pacific herring were deeper in the water column in March 1996 (27.0 - 28.9 m) than in October 1995 (15.0 - 20.2 m) and July 1996 (14.1 - 16.7 m) which were similar (Table 4). Walleye pollock were distributed near the bottom during all three surveys (Table 4). There appeared to be little vertical overlap between Pacific herring and walleye pollock as the mean depths were separated by >29 m of water during all surveys (Table 4).

The proportion of herring schools consisting of a single size cohort varied among seasons. In October 1995,49.0% of the herring schools sampled with the nets consisted of a single size class, primarily age 0 (2 1 schools of the 55 schools sampled). In March 1996, 38.9% of the herring schools sampled with the nets consisted of a single size, primarily age 1 - 2 (14 schools of the 72 schools sampled). In July 1996, 83.3% of the herring schools sampled with the nets consisted of a single size, primarily age 1-2 (21 schools of the 30 schools sampled).

The 26 sites located within the 12 bays had a mean C3NSD value of 3.8 km (SD = 2.45) which was significantly smaller than 9.8 km (SD = 6.69) for the 17 sites located in passages and along open coast line (t-test = -3.6 1, df = 41, p

were similar in length, with a mean of 13.6 mm (SD = 4.52) inside the bays and a mean of 13.4 mm (SD = 3.04) in passages and open coastline (T = 1 1 0 3 , ~ = 0.66).

Pacific herring and walleye pollock C3NSD distributions differed from randomly calculated C3NSD distributions during all three surveys. In October 1995, age 0 and I herring aggregated at the heads of bays with C3NSD values (I1 km) above the mean value calculated from sites outside of the bays (9.8 km; Table 6). Walleye pollock distribution was also significantly aggregated within bays but not as tightly as age 0 and 1 herring (Table 6). In March 1996, all size modes of Pacific herring and walleye pollock had greater than expected C3NSD values ranging between 4 and 10 km (Table 7). In July 1996, age 0 herring aggregated tightly into the heads of bays (57.1 % of C3NSD < 1 km). Age 1 herring were also aggregated at the head of bays (

(Thorne 1983b, Foote 1987, Rose & Leggett 1988, Thorne & Thomas 1990, MacLennan & Simmonds 199 1 , Misund et al. 1995, Huse & Ona 1996, McClatchie et al. 1996, Ona & Mitson 1996, Misund 1997).

Thorne's target strength equation differs from equations calculated for Atlantic herring (MacLennan & Simmonds 1991), however, our estimates of Pacific herring density (means ranging from 0.5 1 to 9.15 fish m-3) were similar to many herring density estimates measured using similar techniques, for example Atlantic herring densities in a fjord in northern Norway ranging from 0.3 to 5.0 fish rn-3 (Misund & Floen 1993) and in the North Sea from 2.94 to 6.68 fish m-3 (Misund et al. 1995). Our density estimates do not agree well with the theoretical density of schooling fish (volume = 0.7 herring body length-)) (Pitcher & Partridge 1979, Pitcher & Parrish 1993). However, in the wild large variations in herring density occur within schools (Misund & Floen 1993) and between schools (Blaxter & Hunter 1982).

The number, density, age and depth of schools of Pacific herring varied seasonally. Pacific herring were tightly aggregated forming fewer, denser schools of a single size cohort near the surface in July. This pattern began to deteriorate in October when schools were slightly less cohesive, and at approximately the same depth. However, the number of schools increased, the number of data cells making up each school decreased and the majority of schools were made up of mixed age cohorts. Pacific herring formed many schools with low densities, and mixed size cohorts, in deeper water (cl fish m3 for herring), in March. The October and March surveys were conducted during the night while the July survey was conducted during daylight. Clupeid distribution is strongly effected on a short temporal scale by the die1 (daylnight) cycle (Blaxter& Hunter 1982, Scott & Scott 1988, Stokesbury & Dadswell 1989, Huse& Ona 1996). However, these observation appear to reflect the conditions in the natural habitat and are therefore comparable to each other because in this northern latitude there is very little daylight in March and practically no darkness during July.

The seasonal difference in the structure of herring schools may result from the availability of food, the physical condition of the fish, and the threat of predation. The inter-fish distance within schools increases with hunger, reducing cohesion, and causing lower mean densities (Robinson 1995). Competition within schools for food is reduced with independent and segregated behavior (Robinson & Pitcher 1989). Fish also increase their chance of encountering food by breaking into small groups or by reducing shoal compactness, thereby increasing the schools foraging area (Robinson 1995). in March herring appeared to be spreating out forming low density aggregations covering large areas. These herring have just survived the winter when prey abundance is minimal and the risk of starvation is high (Paul & Paul 1997, Paul et al. 1997).

The number, density and depths of schools of walleye pollock varied seasonally. Walleye pollock were highly aggregated forming fewer, very dense schools in July and many, low density schools in October and March. Walleye pollock schools were always deep in the water column, near the bottom, and probably less influenced by light conditions. This seasonal variation maybe related to different phases of the life cycle. The walleye pollock collected in July were age 0 and their dense schools seemed to be associated with large jellyfish, Aurelia aurita, aggregations. Large jelly fish, Cvanea capillata, were also frequently seen with several gadids, possibly walleye pollock,

continuously associated with them. The October and March pollock collections were larger juveniles or adult fish, were not associated with jelly fish, and seemed to be pre and post spawning aggregations, respectively.

Pacific herring and walleye pollock distributions were roughly similar, both were bimodal with high concentrations occurring in the northeast and southwest. However, there appeared to be little overlap as they occupied different portions of the water column, Pacific herring in surface waters and walleye pollock near the bottom. Further their morphologies and behavior are different. Therefore competition for prey between these species seems unlikely.

Bays in Prince William Sound appeared to be nursery areas for age 0 and 1 Pacific herring. Larval Pacific herring aggregated within bays in July just prior to metamorphosing into juveniles. Larval herring depend on their transparency to escape detection and capture by visual predators (Batty 1989, Gallego & Heath 1994). The ontogenetic changes from the larval to juvenile body morph (body length 3 1 - 38 mm) require switching to a different antipredator strategy and this is a critical period in terms of vulnerability to predation (Gallego & Heath 1994). As juvenile herring develops its camouflage of guanine reflecting platelets under the scales and black dorsum they also fill their swim bladder and bullae with air (Blaxter 1985). The coupling of the bullae and head lateral line allows herring to detect range as well as direction of sound source and is indispensable to forming and maintaining the schools (Blaxter et al. 1981, Blaxter 1985, Gallego & Heath 1994). Schooling becomes the herring's primary defense against predation for the remaining portion of its life cycle (Blaxter & Hunter 1982).

Age 0 herring were very tightly aggregated near the shore in shallow water at the heads of bays in July. This distribution continued through October but in March the schools had moved away from the shores, into deeper water, however they still remained within the bays. The age 1 herring were also aggregated into tight schools in shallow water near the heads of bays during July and October and dispersed into a scattered layer in March. After their second winter within these bays the juvenile herring appear to join the adult schools leaving the bays at approximately the same time as the new recruits enter the bays (in June- August).

There appears to be a physical difference in the water conditions within bays compared to the open Sound. In October 1995 and July 1996 water temperatures in the upper 30 m were cooler inside these bays than in the open Sound. In March this pattern was reversed and waters were warmer within the bays. What effect this has on juvenile herring metabolic rate is unclear. Juvenile herring require a critical amount of energy to survive the winter (Paul & Paul 1997, Paul et al. 1997) and there may be an advantage to overwintering in cooler waters. The effect these water conditions have on zooplankton production is also unclear. Juvenile herring tightly aggregating near the shore may be feeding on prey associated with the nearshore algae or using the algae as a shelter from predation. Other juvenile species utilize near shore algae in a similar manor (Rangeley & Kramer 1995a, Rangeley & Kramer 1995b). The shift to deeper water away from the shore in March may result from increased food supply being transported into these bays. Adult Pacific herring appear to seasonally migrate within Prince William Sound. A large school of adult herring was observed in Zaikof Bay in March 1996. Large adult schools have traditionally been observed and fished in this bay and a similar large school was observed

here in 1997 (author's unpublished data). These adults moved out of Zaikof Bay and spawned in the Green Island - Stockdale harbor area in April (John Wilcox, personal communitication; Alaska Department of Fish and Game, Cordova). We observed them during the summer months feeding from the lower part of Green Island to the southern part of the Sound, in Latouche and Elrington Passages. There appears to be a great deal of food in this region as the waters leaving Prince William Sound mix with the Gulf of Alaska current. Large schools of Pacific sand lance, large flocks of foraging sea birds and many marine mammals were also observed here during July 1996. The adult herring then migrate back towards Zaikof and are in the Green Island area in late October, as we observed in 1995. The bait fishery traditionally occurs in this area in late October-early November, and occurred in 1996 and 1997 (John Wilcox, personal communitication; Alaska Department of Fish and Game, Cordova).

The predictability in temporal and spatial distributions of herring populations in tidally energetic systems has contributed to overfishing (Sinclair et al. 1985). Further, herring do not exhibit the classic signs of overfishing, i.e. a decline in the catch per unit effort and a loss of larger fish. The dual characteristics of homing to natal spawning areas and larval retention severely restrict the ability of one spawning population to repopulate neighboring spawning areas that have been disrupted by overfishing or a major perturbation (Iles & Sinclair 1982, Sinclair et al. 1985, Sinclair 1988). It is therefore critical to determine the seasonal variability in spatial distributions of each phase of the Pacific herring life cycle in Prince William Sound to determine the potential for recovery of the population and it's role in the ecosystem and effective future management. Acknowlednments We thank R. Foy, M. McEwen, S. Gay and L. Tuttle for assistance with field collections. N. Peters and G. Steinhart assisted in collecting and processing the acoustic data. M. Frandsen assisted with larval fish identification. S. Moreland and M. Vallerino assisted with data analyses. We also thank the captains and crews of the F N Temptation, F N Miss Kayley, F N Kyle David, M N Pacific Star, F N Pagan, R N Pandalus and the Alaska Department of Fish and Game, Cordova, for their help and insights on Prince William Sound. This project was funded by the Exxon Valdez Oil Spill Trustee Council through the Sound Ecosystem Assessment (SEA) project. However the findings presented by the authors are their own and not necessarily the Trustee Council position.

LITERATURE CITED Andersen KP, Ursin E (1978) A multispecies analysis of the effects of variations of effort

upon stock composition of eleven North Sea fish species. ICES Rapp Proc-Verb 172:286-29 1

Auster PJ (1988) A review of the present state of understanding of marine fish communities. Northw At1 Fish Sci 8:67-75

Batschelet E (1981) Circular statistics in biology. Academic Press, London Batty RS (1989) Escape responses of herring larvae to visual stimuli. J Mar Biol Assoc UK

69:647-654 Blaxter JH, Denton EJ, Gray JAB (198 1) Acoustico-lateralis system in clupeoid fishes. In:

Tavolga WN, Popper AN, Fay RR (ed) Hearing and sound communication in fishes. Springer-Verlag, New York, p39-59

Blaxter JHS (1985) The herring: a successful species? Can J Fish Aquat Sci 42 (Suppl. 1):21-30

Blaxter JHS, Hunter JR (1982) The biology of clupeoid fishes. Adv Mar Biol 20: 1-223 Brown ED, Baker TT, Hose JE, Kocan RM, Marty GD, McGurk MD, Norcross BL, Short

J (1996) Injury to the early life history stages of Pacific herring in Prince William Sound after the Exxon Valdez oil spill. Am Fish Soc Symp 18:448-462

Clausen DM (1983) Food of walleye pollock, Theranra chalconramma, in an embayment of southeastern Alaska. Fish Bull 8 1 :637-642

Cushing DH (1967) The grouping of herring populations. J mar biol Ass U K 47: 193-208 Foote KG (1987) Fish target strengths for use in echo intergrator surveys. J Acoust Soc Am

82:98 1-987 Foote KG, Traynor JJ (1988) Comparison of walleye pollock target strength estimates

determined from in situ measurements and calculations based in swimbladder form. J acoust Soc Am 83:9-17

Funk F ( 1 994) Preliminary summary of 1994 sac roe herring fisheries. Alaska Dept. Fish & Game, Juneau

Gallego A, Heath MR (1994) The development of schooling behavior in Atlantic herring Clupea haren~us. J Fish Biol45:569-588 Gunderson RD (1 993) Surveys of Fisheries Resources. John Wilet & Sons, Inc., New York Hansen WR, Eckel EB (1971) Setting and effects of the earthquake. In: The great Alaska

earthquake of 1964. National Academy of Science Hatch SA. Sanger GA (1992) Puffins as samplers of juvenile pollock and other forage fish

in the Gulf of Alaska. Mar Ecol Prog Ser 80: 1 - 14 Hinckley S, Bailey KM, Picquelle SJ, Schumacher JD, Stabeno PJ (1991) Transport,

distribution. and abundance of larval and juvenile walleye pollock (Theranra chalco~ramma) in the western Gulf of Alaska. Can J Fish Aquat Sci 48:9 1-98

Huse I, Ona E (1996) Tilt angle distribution and swimming speed of overwintering Norwegian spring spawning herring. ICES J mar Sci 53:863-873

Iles TD. Sinclair M (1982) Atlantic herring: stock discreteness and abundance. Science 2 15:627-633

Kendall AWJ, Schumacher JD, Kim S (1996) Walleye pollock recruitment in Shelikof Strait: applied fisheries oceanography. Fish Oceanogr 5 (Suppl. 1):4-18

Livingston PA (1993) Importance of predation by groundfish, marine mammals and birds on walleye pollock Theranra chalco~ramma - and Pacific herring Clupea pallasi in the eastern Bering Sea. Mar Ecol Prog Ser 102205-2 15

MacLennan DN, Simmonds EJ (1991) Fisheries Acoustics. Chapman & Hall, London McClatchie S, Alsop J, Coombs RF (1996) A re-evaluation of relationships between fish

size, acoustic frequency, and target strength. ICES J mar Sci 53:780-791 Misund OA (1997) Underwater acoustics in marine fisheries and fisheries research. Rev

Fish Biol Fisher 7: 1-34 Misund OA, Aglen A, Fronaes E (1995) Mapping the shape, size, and density of fish

schools by echo integration and a high-resolution sonar. ICES J mar Sci 52: 11-20 Misund OA, Floen S (1993) Packing density structure of herring schools. ICES mar Sci

Symp 196:26-29 Ona E, Mitson RB (1996) Acoustic sampling and signal processing near the seabed: the

deadzone revisited. ICES J mar Sci 53:677-690 Paine RT, Ruesink JL, Sun A, Soulanille EL, Wonham MJ, Harley CDG, Brumbaugh DR,

Secord DL (1996) Trouble on oiled waters: lessons from the Exxon Valdez oil spill. Annu Rev Ecol Syst 27: 197-235

Paul AJ, Paul JM (1997) Comparisons of whole body energy content of captive fasting age zero Alaskan Pacific herring (Clupea pallasi Valenciennes) and cohorts over- wintering in nature. J Exp Mar Biol Ecol000:000-000 in press

Paul AJ, Paul JM, Brown ED (1997) Fall and spring somatic energy content for Alaskan Pacific herring (Clupea pallasi Valenciennes 1847) relative to age, size and sex. J Exp Mar Biol Ecol000:000-000 in press

Pitcher TJ, Parrish JK (1993) Functions of shoaling behavior in teleosts. In: Pitcher TJ (ed) Behavior of teleost fishes. Chapman & Hall, New York, p363-439

Pitcher TJ, Partridge BL (1979) Fish school density and volume. Mar Biol 54:383-394 Rangeley BW, Kramer DL (1995a) Tidal effects on habitat selection and aggregation by

juvenile pollock Pollachius virens in the rocky intertidal zone. Mar Ecol Prog Ser 126: 19-29

Rangeley BW, Kramer DL (1995b) Use of rocky intertidal habitats by juvenile pollock Pollachius virens. Mar Ecol Prog Ser 126:9-17

Robinson CJ ( 1995) Food competition in a shoal of herring: the role of hunger. Mar Behav Physiol 24:237-242

Robinson CJ, Pitcher TJ (1989) Hunger motivation as a promoter of different behaviors within a shoal of herring: selection for homogeneity in fish shoal? J Fish Biol 35:459-460

Robinson SMC ( 1 988) Early life history characteristics of Pacific herring, Clupea harennus pallasi Valenciennes 1847, in the Strait of Georgia, British Columbia: hydrodynamics, dispersal, and analysis of growth rates. University of British Columbia, Ph.D. thesis, p 14 1

Rose GA, Leggett WC (1988) Hydroacoustic signal classification of fish schools by species. Can J Fish Aquat Sci 45:597-604

Scott WB, Scott MG (1988) Atlantic fishes of Canada. Can Bull Fish Aquat Sci 219: 1-73 1 Sinclair M ( 1 988) Marine populations: an essay on population regulation and speciation.

Washington Sea Grant Program, University of Washington Press, Seattle

Sinclair M, Anthony VC, Iles TD, OBoyle RN (1985) Stock assessment problems in Atlantic herring (Clupea harengus) in the Northwest Atlantic. Can J Fish Aquat Sci 421888-898

Sokal RR, Rohlf FJ (1981) Biometry - the principles and practice of statistics in biological research. Freeman, San Francisco

Spies RB, Rice SD, Wolfe DA, Wright BA (1996) The effects of the Exxon Valdez oil spill on the Alaskan coastal environment. In: Rice SD, Spies RB, Wolfe DA, Wright BA (ed) Proceedings of the Exxon Valdez oil spill symposium. Am Fish Soc Sympos. 18: 1 - 16

Stokesbury KDE, Dadswell MJ (1989) Seaward migration of juveniles of three herring species, Alosa, from an estuary in the Annapolis River, Nova Scotia. Can Field-Nat 103:388-393

Thomas GL (1992) Successes and failures of fisheries acoustics: an international, federal, and regional point of view. Fish Res 14:95- 104

Thorne RE (1977a) Acoustic assessment of hake and herring stocks in Puget Sound, Washington, and southeastern Alaska. In: ARM (ed) Hydroacoustics in fisheries research. ICES Rap et Proc, p265-278

Thorne RE (1977b) A new digital hydroacoustic data processor and some observations on herring in Alaska. J Fish Res Bd Can 34:2288-2294

Thorne RE (1983a) Assessment of population abundance by hydroacoustics. Biol Oceanog 2:253-262

Thorne RE (1983b) Hydroacoustics. In: Nielsen LA, Johnson DL (ed) Fisheries techniques. Am. Fish. Soc., Bethesda MD, p239-259

Thorne RE, Thomas GL ( 1 990) Acoustic observations of gas bubble release by Pacific herring (Clupea harengus pallasi). Can J Fish Aquat Sci 47: 1920- 1928

Thorne RE, Trumble R, Lemberg N, Blankenbeckler D (1983) Hydroacoustic assessment and management of herring fisheries in Washington and southeastern Alaska. 300:269-277

Trumble R, Thorne RE, Lemberg N (1982) The Strait of Georgia herring fishery: a case of timely management aided by hydroacoustic surveys. Fish Bull 80:38 1-388

Wespestad VG, Moksness E (1990) Observations on growth and survival during the early life history of Pacific herring, Clupea pallasi from Bristol Bay, Alaska, in a marine mesocosrn. Fish Bull 88: 19 1-200

Willette M, Sturdevant M, Jewett S (1997) Prey resource partitioning among several species of forage fishes in Prince William Sound, Alaska. In: Forage fishes in marine ecosystems. Alaska sea grant program, AK-SG-97-01: 1 1-29.

Table I . Numbers of dominant fish species collected in Prince William Sound, Alaska in 1995 and 1996. The number of fish measured from seine collections are random subsamples (1000 individuals) of the total catch therefore the % reflect the subsample maximum.

October 1995 March 1996 July 1996 trawl seine trawl seine seine

scientific name common name 74 72 67 5 9 60 % Clupea pallasi Pacific herring 1421 47803 11535 40588 22639 65.0 Theranra walleye pollock 3308 13994 7319 3950 7929 19.2 chalcogramma Ammodytes hexapterus Pacific 1 1 5000 2.6

sandlance Mallotus villosus capelin 68 266 3406 2.0

Table 2. Clupea pallasi and Theragra chalcogramma. Chi-squared tests examining the polar angle distributions of the three size modes of Pacific herring and walleye pollock during three acoustic surveys of Prince William Sound, Alaska in 1995 and 1996; NS = no significant difference, p>0.05.

October 1995 March 1996 July 1996 df X 2 p = df X 2 p = df X 2 p =

herring age 0, 1 and adults 8 109 0.01 7 41.5 0.01 6 87.7 0.01 herring age 0 and 1 8 0.9 NS 7 20.9 0.01 6 18.2 0.05 herring age 0 and adults 8 85.0 0.01 7 40.1 0.01 6 87.9 0.01 herring age 1 and adults 8 80.8 0.01 7 9.7 NS 6 43.4 0.01 herring and poIlock 8 227 0.01 7 65.8 0.01 6 240 0.01 herring age 0 and pollock 8 1 17 0.01 7 9.0 NS 4 90.2 0.01 herring age 1 and pollock 8 114 0.01 7 23.2 0.01 4 62.2 0.01 herring adults and pollock 7 29.6 0.01 7 31.7 0.01 6 67.1 0.01

Table 3 Clupea pallasi and Theragra chalcogramma. Estimated means and standard deviations (SD) of densities (fish m-3) and number of data cells for the three size modes of Pacific herring and walleye pollock aggregations observed during three acoustic surveys of Prince William Sound, Alaska in 1995 and 1996 (n = count).

October 1995 fishm-' March 1996 f i ~ h . m - ~ July 1996 fish.m-)

herring age 0 age 1 adult

polloc k herring

age 0 age 1 adult

polloc

n mean SD 137 2.52 8.36 126 0.51 1.10 54 1.35 3.21

119 3.14 8.68

cells (40- 120 x 1 m) 137 15.55 39.58 126 13.77 37.42 54 9.63 14.04

1 19 8.05 1 1.07

n mean SD 179 0.91 2.84 223 0.81 0.95 147 0.51 0.95 148 4.20 14.59

cells (40 x 1 m) 179 22.50 65.1 1 223 25.00 82.82 147 28.36 96.88 148 13.12 44.73

n mean SD 8 2.67 5.46

42 9.15 15.75 32 3.62 0.86 31 267.87 72.05

cells (20 x lm) 8 98.25 155.28

42 31.55 85.97 32 21.00 56.98 31 300.16 72.05

Table 4 Clupea pallasi and Theragra chalcogramma. Estimated means and standard deviations (SD) of three size modes of Pacific herring and walleye pollock depth in the water column and sea bed depth during three acoustic surveys of Prince William Sound, Alaska in 1995 and 1996 (n = count).

depth of fish (m) October 1995 March 1996 July 1996

herring n mean SD n mean SD n mean SD age0 2131 15.02 6.51 3675 27.48 15.82 780 14.14 9.38 age 1 1819 15.48 6.65 5463 27.04 13.83 1325 15.18 10.42 adult 520 20.24 7.72 4180 28.86 13.66 633 16.65 10.58

polloc 958 64.44 26.42 1942 57.88 25.27 6301 63.75 34.65 k herring bottom depth (m)

age 0 2131 47.87 26.01 3420 47.08 20.69 723 34.49 21.62 age 1 1819 45.62 26.10 5016 47.46 19.46 1242 34.96 18.28 adult 520 45.96 24.95 4026 46.43 16.76 583 47.42 38.35

polloc 958 68.7 30.96 1872 63.09 22.46 5612 86.99 34.65 k

Table 5 Mean and standard deviations (SD) of water columns inside bays and outside bays in passages and along open coastline during three surveys of Prince William Sound, Alaska in 1995 and 1996 (n = count). The Mann-Whitney rank sum test was used to determine differences between mean values (T).

water depth (m) temperature (oC) salinity (oloo) October 1995 n mean SD mean SD

0-30 inside 146 9.34 0.626 T=16135 28.77 1.111 T=18583 0-30 outside 180 9.66 0.821 p

Table 6. C l u ~ e a pallasi and Thera~ra chalcogramma. Chi-squared analysis comparing the percent frequency distributions of observed sum of the three near shore distances (C3NSD) for Pacific herring and walleye pollock aggregations to values generated from a random distribution, during the October 1995 acoustic survey in Prince William Sound, Alaska; * ~ ~ 0 . 0 5 , ** p ~ 0 . 0 1 .

Pacific herring (%) walleye C3NSD (km) age 0 age 1+ adult pollock (%) random (%)

1, 2 35.38 32.77 4.08 19.82 15.59

Table 7. Clupea pallasi and Theragra chalcogramma. Chi-squared analysis comparing the percent frequency distributions of observed sum of the three near shore distances (C3NSD) for Pacific herring and walleye pollock aggregations to values generated from a random distribution, during the March 1996 acoustic survey in Prince William Sound, Alaska; * pc0.05, ** p ~0 .01 .

Pacific herring (%) walleye C3NSD (km) age 0 age 1+ adult pollock (%) random (%)

1 7.74 8 .OO 3.10 3.13 11.18 2 22.58 20.00 17.83 15.63 26.7 1 3 18.71 18.86 17.83 21.09 18.01 4 14.84 15.43 16.28 14.06 9.32

5, 6 16.13 15.43 18.60 14.84 6.2 1 7 to10 15.48 17.14 19.38 18.75 5.59

11,12,13 0.65 1.7 1 2.33 6.25 6.21 14 to 17 2.58 2.29 3.10 3.13 6.2 1 18 to 29 1.29 1.14 1.55 3.13 6.83

>30 0.00 0.00 0.00 0.00 3.73 X2= 53.68** 58.40** 84.55** 63.60** df= 9 9 9 9

Table 8 Clupea pallasi and Theragra chalco~ramma. Chi-squared analysis comparing the percent frequency distributions of observed sum of the three near shore distances (C3NSD) for Pacific herring and walleye pollock aggregations to values generated from a random distribution, during the July 1996 acoustic survey in Prince William Sound, Alaska; * p

List of Figures.

I. Location of the October 1995 (solid line), March (dashed line) and July 1996 (dotted line) acoustic survey transects in Prince William Sound, Alaska; EP = Elrington Passage, LP = Latouche Passage, GI = Green Island, SH = Stockdale Harbor, ZB = Zaikof Bay; + was positioned at 60' 60.00' N 146' 90.00' W, 0 = true north.

2. Clupea pallasi. Percent size frequency distributions of Pacific herring (fork length mm) collected during three acoustic surveys of Prince William Sound, Alaska, in 1995 and 1996 (n = count).

3. Theranra chalconramma. Percent size frequency distributions of walleye pollock (fork length mm) collected during three acoustic surveys of Prince William Sound, Alaska, in 1995 and 1996 (n = count).

4. Clupea pallasi and Theragra chalco~ramma. Polar angle percent frequency distributions of Pacific herring and walleye pollock observed during three acoustic surveys of Prince William Sound, Alaska, in 1995 and 1996.

Fig. 1. Stokesbury et al.

October 95; n = 7626

March 96; n = 33655

15 1

July 96; n = 22700

fork length (mm)

Fig. 2 Stokesbury et al.

October 95; n = 1487

March 96; n = 9561

July 96; n = 1864

fork length (mm)

Fig. 3 Stokesbury et al.

October 1995 age 0 herring 69 age 1 + herring

60 adult herring

8 40

20

0

March 1996 age 0 herring N age 1 + herring a d u l t herring

pollock

July 1996 age 0 herring 80 N age 1 + herring 60 adult herring

8 40

20

0

Degrees

Fig. 4 Stokesbury et al.

Appendix I1

Preliminary Documentation of Temporal and Spatial Variability of Pacific Herring, Other Forage Fish, and Seabirds in Prince William Sound, Alaska

Evelyn D. Brown April 7, 1998

Appendix 11. Preliminary Documentation of Temporal and Spatial Variability of Pacific Herring, Other Forage Fish, and Seabirds in Prince William Sound, Alaska

Evelyn D. Brown Introduction

This report covers preliminary aerial survey data presented at the annual science reviews for both the Sound Ecosystem Assessment (SEA) and Alaska Predator Ecosystem Experiment (APEX) projects funded by the Exron Valdez Oil Spill Trustee Council (EVOS TC). From 1995 through 1997, monthly broadscale aerial surveys were flown in Prince William Sound (PWS) and the Outer Kenai (OK) over 'a 5-7 day period during May through August (see Appendix III and Stokesbury et al., 1997). Peak counts of fish schools were recorded in June and July. Less schools were observed in May and August. Increased counts observed in June and July were probably due to a change in fish distribution to shallow surface waters (see Stokesbury et. al., Appendix I this report) as the aerial survey is limited to observations of surface schooling fishes. Also, by July metamorphosis of larval Pacific herring (Clupea pallasi; Stokesbury et al., Appendix I) and sandlance (Arnmodvtes hexapterus) occurs and schooling behavior begins (evidenced by reductions in larval abundance, Norcross et al. 1996 and observations of large numbers of age-0 juvenile sandlance in the nearshore, APEX, unpublished data). Finally, pre-spawning capelin (Mallorus villosus) and post-spawning eulachon (Thaleichthys pacificus) form large visible schools in June in PWS and adjacent waters of the Gulf of Alaska (E. Brown, unpublished data).

I will present here the graphic results of the broadscale surveys during the months of June and July for PWS.

Methods

Methodology for the aerial survey technique is documented by Brown and Norcross (in prep) and in Appendix III of this report. For this report, the aerial survey database was queried for the broadscale survey data from June and July. Repeat surveys over herring nursery bays of interest to the SEA project were left out of the query. The data therefore represents a single pass over the entire area over the course of a week. The only shorelines not sampled were in the fiords of the extreme northwest corner of PWS (Port Wells, not labeled on the maps). Surface areas were estimated for all schools and totai surface area estimates were plotted for each unique latitude and longitude by species (there were often several schools at a given location). The results were plotted using the GMT mapping software and the key for the shade-by-value school plots is given in Figure 1 .

Seabird behavior codes are recorded easily from the air including plunging, milling tightly aggregated on the water, resting on the water (loose aggregation), resting on shore, broad- area search, and travelling. Therefore, a determination of the numbers of birds foraging (plunging, milling, tightly aggregated on the water) versus searching or other behaviors can be made. I can determine not only the degree to which birds are associated with surface

schools, but also what their behaviors are in association with schools. Therefore, the degree of spatial and temporal variability in foraging and school association can be determined.

The aerial database includes mainly white gulls (black-legged kittiwakes, mew gulls, and glaucous-winged gulls) which are easy to sight from the air. They also do not spend any time under the surface. Diving ducks and other seabirds are difficult to sight from the air. Therefore, the seabirds represented are gull species.

Results and Discussion

All forage species and seabirds or gulls were plotted by month and year (Figures 2-7). The notable features are discussed. The existence of large numbers of pre-spawning capelin schools were observed in June of 1995, but not in subsequent years (Figures 2-4). These sightings were validated by targeted catches on pre-spawning capelin sighted from the air. In June of 1997, following reports of excessive interceptions of spawning eulachon by Copper River drift net salmon fishermen (Copper River Fishermen's United, Cordova, Alaska, personal communication), large numbers of huge eulachon schools (about 1000 m') were sighted mainly off Montague Island (Figure 4). These sightings were validated by examinations of stomach contents of predatory fish caught near the school groups and by interceptions of eulachon by salmon fishermen. No eulachon were observed in June of 1995 and 1996. Large numbers of juvenile herring schools were sighted all three years in June representing mainly age- 1 herring (see Appendix El). However, interannual variability in abundance was evident with peak surface area estimates occurring in 1996 (Figures 2-4). This could represent differences in overall year-class strength. Only in June of 1997 were sandlance schools observed in significant numbers. Surface water temperatures were abnormally high in 1997 (SEA, unpublished data) and the anomalous numbers of sandlance sighted in June could represent early recruitment of post- metamorphic sandlance to nearshore beaches.

In June, broadscale seabird foraging activities were focused on the pre-spawning capelin and in eastern PWS in 1995 (Figure 2) and post-spawning eulachon in 1997 (Figure 4). In general, more foraging activity was observed in 1997 over the two previous years. In 1996, bird foraging activities were more numerous at nearshore beaches on outer Montague and western and northern Hinchinbrook Island. There was also large number of foraging flocks in northwestern PWS near the village of Tatitlek and associated with juvenile herring schools.

The notable features in July of each year were the absence of capelin and eulachon and peak annual counts of both juvenile herring and sandlance (except in 1995 when peak herring counts occurred in June) (Figures 5-7). The variability of herring counts in July probably represents changes in year-class strength for both age- 1 and age-0 herring. However, changes in the depth distribution of juvenile herring as well as early departure of age- I + herring from nursery bays could also explain the variability observed (see Stokesbury et al., Appendix I). Comparisons of acoustic and aerial data, planned for later

this spring, will document variability in distribution by depth, which will aide in correct interpretation of the aerial results (see Appendix III). No such comparisons are available for sandlance since they generally occur in waters too shallow for acoustic measurements (Appendix III). However, because of that characteristic, aerial counts are probably more conclusive for sandlance than for juvenile herring at this time. Sandlance appear to be in a state of population building as increasingly higher numbers of schools were observed from 1995 to 1997 (Figures 5-7).

Although there was considerably bird foraging activity associated with schools sighted from the air in July, flock size appeared to have decreased from June to July (compare the size of the triangles from Figures 2-4 to 5-7). There appears to be a species preference by the foraging seabirds for juvenile herring over sandlance. In 1996 and 1997, 35.7% (n=226) and 43.6% (n=326) respectively of the herrings schools were associated with seabirds. This compares to 18.3% (n=7 1) and 1 1.1 % (n=180) of the sandlance schools in 1996 and 1997 respectively. There did not appear to be a preference for school size as the frequency distribution of school size was identical for schools with or without birds.

In all three years, the majority of the seabirds sighted from the air were involved in active foraging behaviors (41.0%) or search activities !27.0%). The remainder were either resting or travelling ( 1 5.7%) including counts at seabird colonies (mainly kittiwakes).

It appears from preliminary results of APEX chick diet studies, that PWS seabirds feeding primarily on pelagic fishes are targeting juvenile herring and sandlance. Since sandlance and juvenile herring are the most abundance surface schooling fishes observed in PWS, the birds appear to be feeding on what is available. I produced plots to show interannual variability of juvenile herring and sandlance in the months of June and July. The peak abundance estimates (total school surface areas) of juvenile herring (probably age- I ) in June of 1996 are obvious in Figure 8. For sandlance, the 1997 peak June count is obvious (Figure 9). In July, both peak school counts and surface area estimates in 1996 and 1997 exceeded those in 1995 (Figure 10). There is also evidence of slight changes in distribution from year to year. In 1997, more herring schools were observed in waters outside bays than in 1996. I have no explanation for this. The population increase in sandlance is obvious in Figure 11 with an accompanying increase in spatial distribution from 1995 to 1997.

Now that the documentation of fish distribution is complete, I can proceed with ecological evaluations of those distributions. I can compare the distributions to oceanographic conditions and regimes occurring within and outside of PWS. I can compare the trends in abundance of the juvenile herring to resulting changes in the adult population size and distribution. 1 can examine regional and temporal changes in seabird or gull foraging activity as well as interannual fluctuations in colony size (on a broadscale). In short, the potential for use is broad-based. There are obvious monitoring advantages in the continuation of this data set over a longer period of time.

Literature Cited

Brown, E.D. and B.L Norcross. In prep. Assessment of forage fish distribution and abundance using aerial surveys: survey design and methodology. To be submitted to Ecological Applications.

Appendix I t Figure I . Key for following figures.

Shade by value of fish schools at Total no. birds at a given location, species by color a given location

Herring 0 Eulachon 0 Sandlance

0 Capelin A Birds

Appendix U Fig. 2. Distribution of Seabird and Forage Fish. in Prince William Sound.

Alaska, June 1995. No eulachon were sighted.

Herring 0 Eulachon 0 Sandlance

Capelin A Birds

Appendix I1 Fig. 3. Distribution of Seabird and Forage Fish. in Prince William Sound,

Alaska. June 1996. No eulachon were sighted.

Herring Q Eulachon 0 Sandlance

a Capelin A Birds

Appendix 11 Fig. 4. Distribution of Seabird and Forage Fish, in Prince William Sound.

Alaska. June 1997. No capelin were sighted.

Herring 0 Eulachon 0 Sandlance

0 Capelin Birds

Appendix II Fig. 5. Distribution of Seabird and Forage Fish, in Prince William Sound.

Alaska, July 1995. No eulachon were sighted.

Herring 0 Eulachon 0 Sandlance

a Capelin A Birds

Appendix I1 Fig. 6. Distribution of Seabird and Forage Fish, in Prince William Sound.

Alaska, July 1996. No capelin, or eulachon were sighted.

Herring

* Capelin 0 Eulachon 0 Sandlance

A Birds

Appendix I1 Fig. 7. Distribution of Seabird and Forage Fish, in Prince William Sound.

Alaska, July 1997. No capelin were sighted.

10-55

Appendix U Fig. 8. Interannual Variation in the Distribution of Herring,

in Prince William Sound, Alaska, June 1995-1997.

Appendix 11 Fig. 9. Interannual Variation in the Distribution of Sandlance, in Prince

William Sound. Alaska, June 1995- 1997.

Appendix 11 Fig. 10. Interannual Variation in the Distribution of Herring, in Prince

William Sound, Alaska. July 1995- 1997.

Appendix I1 Fig. 1 1. Interannual Variation in the Distribution of Sandlance, in Prince

William Sound, Alaska, July 1995- 1997.

Appendix I11

Pro